Towards Visuospatial Cognition via Hierarchical Fusion of Visual Experts

TL;DR

This paper introduces ViCA2, a multimodal large language model designed specifically for visuospatial reasoning, utilizing a hierarchical fusion architecture and a large spatial question-answer dataset, achieving state-of-the-art results.

Contribution

The paper presents ViCA2, a novel MLLM with hierarchical fusion of visual experts and a new large-scale spatial dataset, advancing visuospatial cognition capabilities.

Findings

ViCA2-7B achieves 56.8 on VSI-Bench, outperforming larger models.

Hierarchical fusion improves spatial reasoning accuracy.

The dataset ViCA-322K enables targeted instruction tuning.

Abstract

While Multimodal Large Language Models (MLLMs) excel at general vision-language tasks, visuospatial cognition - reasoning about spatial layouts, relations, and dynamics - remains a significant challenge. Existing models often lack the necessary architectural components and specialized training data for fine-grained spatial understanding. We introduce ViCA2 (Visuospatial Cognitive Assistant 2), a novel MLLM designed to enhance spatial reasoning. ViCA2 features a dual vision encoder architecture integrating SigLIP for semantics and Hiera for spatial structure, coupled with a token ratio control mechanism for efficiency. We also developed ViCA-322K, a new large-scale dataset with over 322,000 spatially grounded question-answer pairs for targeted instruction tuning. On the challenging VSI-Bench benchmark, our ViCA2-7B model achieves a state-of-the-art average score of 56.8, significantly surpassing larger open-source models (e.g., LLaVA-NeXT-Video-72B, 40.9) and leading proprietary models (Gemini-1.5 Pro, 45.4). This demonstrates the effectiveness of our approach in achieving strong visuospatial intelligence with a compact model. We release ViCA2, its codebase, and the ViCA-322K dataset to facilitate further research.